The present disclosure may be used in connection with essentially any wireless communication system. However, for pedagogical and exemplary reasons, the disclosure is hereafter described in relation to a 3rd Generation Partnership Project (3GPP) Long Term Evolution (LTE) system.
First, a basic description of a 3GPP LTE system is given. As shown in FIG. 1, the architecture of 3GPP LTE is simplified compared to UTRAN release 6, by removing the central node controlling multiple Node Bs (RNC) and instead simply defining two nodes, E-UTRAN Node B (eNB) and access Gateways (aGWs), where the eNB belongs to the evolved UTRAN (E-UTRAN) and the aGW belongs to the evolved packet core network (EPC). As also shown in FIG. 1, there exists an X2 interface between the eNBs and an S1 interface between the eNB and aGW. This is described in reference document: 3GPP TS 36.300, v0.2.0, “Evolved Universal Terrestrial Radio Access (E-UTRAN) and Evolved Universal Terrestrial Radio Access Network (E-UTRAN) (Release 8)”, 2006-11.
FIG. 2 shows the working assumption on the control plane (CP) architecture for the new LTE Radio protocol stack in 3GPP. The Radio Link Control (RLC) and Medium Access Control (MAC) protocols are terminated in the User Equipment (UE) and in the E-UTRAN Node B (eNB), while the header compression and security responsible protocol, named Packet Data Convergence Protocol (PDCP), are terminated in the UE and the access Gateway (aGW).
An intra E-UTRAN handover, where a UE is handed over from a source eNB to a target eNB, in an RRC_CONNECTED state (defined below) of a background art system, is a UE assisted network controlled handover, with handover preparation signaling in E-UTRAN. FIG. 3 shows a basic handover scenario where core network nodes do not change.
At a handover preparation procedure, context data is transferred from source eNB to target eNB. The context data, earlier discussed in 3GPP, is primarily related to the non-dynamic RRC (Radio Resource Control) configuration, i.e., the parameters that were received from the core network when radio bearers were setup and all the RRC configuration information (e.g., radio bearer and security RRC configuration).
The notation control channel is hereafter used in this document for control channels or control mechanisms of the wireless interface between the UE and the Node B, e.g., control channels or control mechanisms of the wireless Layer 1 or Layer 2. These control channels or control mechanisms control, for example, modulation and coding or retransmissions or reporting of frequently needed measurement data, such as channel quality or UL buffer status in the UE. Such a control channel may be a separate channel or multiplexed in-band with other channel(s) or information
In 3GPP LTE, Non-Access Stratum (NAS) protocol states and state transitions are defined. There are three UE specific LTE states, which have been agreed in the standardization work and described in reference document: 3GPP TS 36.300, v0.2.0, “Evolved Universal Terrestrial Radio Access (E-UTRAN) and Evolved Universal Terrestrial Radio Access Network (E-UTRAN) (Release 8)”, 2006-11. These three UE specific LTE states so far have been denoted as LTE_DETACHED, LTE_IDLE, and LTE_ACTIVE. The agreed LTE states capture the mobility management and are defined in the following way:                LTE_DETACHED state: The location of the UE is not known by the network (e.g., the UE power is switched off).        LTE_IDLE/RRC_IDLE state: The UE location is known on a tracking area in aGW. The UE can be paged and the UE registers to the network on tracking area change. UE performs cell reselection during mobility. The RAN/eNB does not maintain a UE context.        LTE_ACTIVE/RRC_CONNECTED state: The UE location is known on a cell level. The network directs the UE to serving cells. The RAN/eNB does maintain a UE context.        
Further, there is, in reference document: 3GPP TR 25.913, v7.0.0, “Requirements for Evolved UTRA (E-UTRA) and Evolved UTRAN (E-UTRAN) (Release 7)”, 2005-06, requirements that include a “Dormant” MAC-substate within the LTE_Active state used for UE battery preservation and radio resource preservation. The MAC state is here denoted MAC_Dormant, to be distinguished from the active MAC state denoted MAC_Active. A LTE state model, including the MAC-Dormant state, is shown in FIG. 4. The main characteristics of the states within LTE_Active are:                MAC_Dormant: Power and resource preservation may be achieved through discontinuous transmission and reception in the UE.        MAC_Active: The UE is prepared for direct transmission (Uplink) and reception (Downlink).        
The system described so far is still under discussion in the 3GPP standardization work. It is here given as an example of how a system may be implemented. Details in the system might be changed in the continuous future discussions during the standardization work, such as names of definitions for states or the like in the system. A skilled person, however, realizes that the present disclosure may be implemented in a system having such possible differences from the above described system, as well as in essentially any other wireless communication system.
LTE is a packet oriented system, in which packet arrivals are in general bursty and discontinuous so that battery saving from Discontinuous Reception (DRX) operation could be significant. It is also anticipated that UE will stay in the connected mode relatively long in a LTE system, so that efficient DRX operation for connected mode UE is important. FIG. 5 shows a DRX structure proposed in reference document: Samsung, R2-062778, “DRX operations for connected mode UEs in LTE”, 2006-10, where “Active Period” is the period during which a UE's transmitter/receiver is turned on, “Sleep Period” is the period during which a UE's transmitter/receiver is turned off and “DRX cycle length” is the time distance between consecutive active period start positions. The DRX related parameter configuration could depend on the LTE states, service characteristics, radio environment, UE capability, or the like.
A timer in Node B could control the switch from MAC_Active to MAC_Dormant. Alternatively, the switch could be based on a UE indication, such as a measurement, but then the timer would need to be transmitted to UE from the eNB. For even more fine-grained control, using multiple DRX cycle lengths in a MAC_Dormant state, multiple inactivity timers could be used. This has been proposed in the 3GPP standardization work. This solution is based on the assumption that activity is assumed higher during some period after a data burst for normal traffic patterns.
The DRX setting would then be determined by:                Configuration Information (RRC); and/or        Activity (or inactivity) history.        
As the LTE system is a packet oriented system, switching from the LTE_Active to the LTE_Idle state could also be based on inactivity timers, rather than based on the setting up and release of connections.
An Uplink, from UE to eNB, in the LTE radio access interface is orthogonal, meaning that different uplink transmission resources are isolated from each other. Furthermore, the mechanism for resource isolation is such that allocation of a first resource implies that the available resources for other usage are decreased, even if the first resource is not used for an actual transmission. This means that it is a waste of resources to allocate dedicated transmission resources to a UE if the UE does not use these resources.
A normal condition for starting uplink data transmission for bulk data transfer in LTE is that a UE from the start does not have any dedicated uplink transmission resources. The UE then has to perform a “random access” procedure on a common contention channel, in order to be allocated an uplink transmission resource by the eNB, which owns the transmission resources. When being in data transmission phase, the UE can request more uplink resources by inband signaling, by indicating the need to send more data in a MAC header. The signaling, from UE to eNB, to request uplink resources by a random access procedure, and the signaling, from eNB to UE to grant certain uplink transmission resources can involve significant overhead. For low bandwidth media applications, such as video or voice, the overhead of the above mentioned uplink resource allocation procedures may be very significant, resulting in a problem relating to efficient usage of communication resources in the system.
This problem has been recognized in 3GPP and various approaches have been proposed to resolve this problem. Most proposals presented in 3GPP discussions involve some kind of “persistent” scheduling, meaning that the eNB does not allocate resources for every data burst, but instead determines an allocation of time-recurring resources needed for transmitting a stream of certain bandwidth. That is, in “persistent” scheduling resources for a number of following data bursts are allocated at once, which may lower the resource allocation signaling overhead. Further, in order to be able to utilize non-activity, e.g., for voice, the eNB could change, or remove, the resource allocation when inactivity has been detected.
For some IP Multimedia Subsystem (IMS) media applications, there is a possibility for the Core Network to send Quality of Service (QoS) related parameters to an eNB in the Radio Access Network (RAN), so the eNB can deduce from those parameters how this “persistent” resource allocation shall be configured to match the application needs. For instance, different application codecs might generate different sized data packets at different periodicity, etc. The Core Network may for IMS provide these parameters, since IMS uses standardized traffic procedures. The IMS Network can therefore easily from the IMS signaling extract the needed parameters and provide them to the eNB, using the Core Network.
However, for various Internet applications (other than IMS), there will in many cases be no such QoS information available, since the Core Network does not from the start know the characteristics of the received signals as this large amount of various Internet applications are not as strictly regulated as the IMS applications. The Core Network can therefore from the start of the application not inform the eNB how to configure the resource allocation of the communication. Furthermore, a radio interface ciphering that is terminated in the Core Network would make it difficult for the Radio Access Network or the Node B to quickly detect codec type, etc., based on peeking into the data stream.
In background art solutions of this problem, the eNB would then start by over-allocating uplink resources. Based on UE usage and on UE in-band requests for more resources or absence of such requests, the eNB could then eventually learn the codec behavior, i.e., it could eventually have learned the used Packet Data Unit (PDU) sizes, the packet periodicity and the activity behavior. This method, however, does not fully optimize the resource allocation and usage, since the eNB has to over-allocate resources during the time period of learning the codec behavior.
Also for IMS applications, the eNB is not always provided with all parameters needed. For instance, the eNB is not always provided with information regarding packet periodicity. Therefore, also for some cases relating to IMS application, a learning procedure has to be performed by the eNB, during which the resource allocation will be non-optimized.
There is thus a resource allocation problem present in the background art solutions, resulting from the fact that the eNBs do not have all the adequate information it needs for configuration of the resource allocation and therefore has to start by, during a time period, learning the behavior of the application codec. This may be especially troublesome in mobility scenarios, when a UE is moving in the system, since the UE then is handed over from eNB to eNB, and each of these eNBs has to carry out this learning procedure.